Introduction: We describe a new approach to drug discovery which joins
the technologies of medicinal and combinatorial chemistry, allowing s
election of the most active variant of a lead compound from a large (>
10(12)) Pool. A small-molecule covalent inhibitor of elastase was coup
led to a randomized pool of RNA, and this assembly was iteratively sel
ected for oligonucleotide sequences that promote the covalent reaction
of the inhibitor with the human neutrophil elastase (hNE) active site
. Results: Incorporation of the covalent inhibitor into the randomized
pool increases the second-order rate of inactivation of hNE by simila
r to 15-fold; sequences selected from this pool show an additional sim
ilar to 20-fold increase in activity The relative rate of cross-reacti
on with another serine protease, cathepsin G, was reduced >100-fold. L
ow doses of the inhibitor were found to prevent lung damage inflicted
by human neutrophils in an isolated rat lung model of acute: respirato
ry distress syndrome (ARDS). Conclusions: This result supports the hyp
othesis that neutrophil elastase is a significant effector of inflamma
tory disease. More generally, our findings demonstrate that blending s
mall molecules into combinatorial libraries is a feasible method of dr
ug discovery.